16 Nov 2015 13:36:06
Hi Eds,
Currently doing a statistics project where we fit a model which will say where a club will finish in the league (only doing a few teams like Liverpool, Utd, Arsenal, City, Everton, spurs and villa) . I've been looking to find data for excel spresdsheets on totals spent on transfers each year and various other stats but have no joy so far. Was wondering if yee would have any idea where I could find this data? (Looking back maybe 10-15 years or so)
Cheers.
{Ed002's Note - The accounts of all clubs are available and will give you the annual expenditure on wages, transfers and, critically, other payments such as agent fees (which you cannot exclude - as an example, Liverpool have spent £40M plus VAT on agent fees over the past three years). If you use transfer fees alone (which Transfermarkt will summarise and have correct at the start of the season but not during the season (due to the way the obtain data) then you will have a distorted view. Finding someone with access to the UEFA data will give you the correct transfer fees but not any of the additional data.
Aside from that I don't know what other data you are unable to find. f you want any such model to be valid, you need to build in such variables as injuries.}
16 Nov 2015 16:19:12
it sounds a big task of a job mate good luck in it.
16 Nov 2015 18:06:24
Not sure whether simple statistic model like logictic regression will give you an answer in your predictions, seems to be more complex econometry problem, but good luck.
16 Nov 2015 21:54:23
Agree with eomer, would be really interested to see the result!
Can't guarantee how accurate the prices are but seem to be to what was reported. Probably a good starter for 10.
http://www.transferleague.co.uk
Good luck!
Z.
17 Nov 2015 01:39:54
Cheers Ed002, luckily for me the lecturer is an idiot and knows nothing about football. Doing only a basic enough regression model to predict the finishing position of a premier league team based on their net transfers, wages and other data with about 5 or 6 predictor variables. He really just wants us to be able to use the statistical software for real world problems so it wouldn't have to be ground-breakingly accurate! Cheers for the info!
{Ed002's Note - If that is the criteria, it won't be accurate at all I am afraid.}
17 Nov 2015 23:14:35
Oh no I know full well it won't be accurate but its something easy to talk about for myself and he won't have a notion of what's going on anyway so it suits all parties involved.
Cheers.